SummaryNo relevant deficiency of TFPI or genetic polymorphisms could thus far consistently be associated with venous thromboembolism. We hypothesized that the substrates of the TFPI protein, including FVII or FX (rather than the protein itself) could induce a hypercoagulable state. We created a novel TF-based clotting assay that evaluated the anticoagulant response to exogenously added recombinant TFPI. The response to TFPI was expressed as the ratio of the clotting time with and without TFPI. By using 118 healthy controls, we established a reference range between 1.31 and 1.93 (mean value ± 2 standard deviations (SD), 1.62 ± 0.31). We then evaluated samples from 120 patients with a history of venous thromboembolism but no evidence of hereditary and acquired thrombophilia. The range of the patients’ ratios was significantly (P < 0.001) lower, falling between 1.2 and 1.78 (mean value ± 2 SD, 1.49 ± 0.29). Of the 120 patients, 39 (32.5%) had a TFPI sensitivity ratio below the 10th percentile of the controls, compared with 11 (9.3%) of the healthy controls. The crude odds ratio for venous thrombosis for subjects with a TFPI sensitivity ratio below the 10th percentile was 13 (95% CI; range, 3.1 to 54.9) compared with those with a ratio above 1.8 (90th percentile). Patients with idiopathic thromboembolism did not have a decreased TFPI sensitivity ratio more often than patients with thrombosis with a circumstantial risk factor. Based on these results, a reduced response to TFPI may lead to an increased risk of venous thrombosis.